Here are my qualifications: I've been using UNIX for 32 years, starting with ed, the original line editor. No screen editors then because we didn't have screens. They didn't show up until around 1975. Soon after Bell Labs employees who went to Berkeley started bringing vi back with them.
Anyway, now for my opinions:
1. One must distinguish between the learnability of an editor and its usability. (True of any user interface, in fact.) For UNIX character-oriented screen editors, there is a huge difference between the two. For other user interfaces the difference is not so great, but it is in this case. Go to a beginner learning vi and you will find that the modes are confusing. Go to an experienced vi user, and you will find that he or she is barely aware of the modes.
2. I edit UNIX files a lot, but in two very different ways: The first is occasional use, right at the console, when I'm unfamiliar with the system or just setting it up. In this case the last thing I want to do is install an editor. I would be happy to use ed; I just want something quick to, for example, set up an fstab file. The second kind of use is for everyday editing, and for that I'm willing to take the time to install what I want. Actually, though, because I have so many different machines, I don't bother. As soon as the system is able to, I go to my primary machine and access the new machine via SSH or Xterm.
I used to think, as many others here do, that vi was the essential editor to learn because it is always there. Imagine my surprise when I installed gentoo the other day and found no vi! Something I'd never heard of, nano, was there instead. OK, fine, Control-o to write the file. At least it was listed at the bottom of the screen. Soon I "emerged" vi (vim, really), and the Gentoo system was back in the fold.
For everyday use, I never use a character editor anyway. I did once, but stopped around 10 or 12 years ago. (Guys, I'm way older than most of you, but even I know that it's 2004!) So for me (given my description of the roles editors play, above) joe has no use whatsoever. In fact, it solves a problem that I don't actually have.
Nonetheless, I wish the new joe project all the best. WordStar (which I used a bit in the mid-1980s) was a great piece of work, and it's nice to see it live on.